Direct Detection of Planets Mark Clampin GSFC. Mark Clampin/GSFC Introduction Definition –Direct detection of extrasolar planets (ESPs) by imaging –Nobody.

Direct Detection of PlanetsDirect Detection of Planets

Mark ClampinMark ClampinGSFCGSFC

Mark Clampin/GSFC

Introduction

• Definition– Direct detection of extrasolar planets (ESPs)

by imaging

– Nobody has yet imaged an extrasolar planet!

– Goals of talk constrained by recent HST devlopments

– Discuss what might be done with existing instruments

• ACS, NICMOS, STIS

• Planet detection is a key theme of Origins Roadmap– HST has a role to play

Mark Clampin/GSFC

History

• Detection of extra-solar planets was always a science goal for HST – However, requirement were not well understood– Brown and Burrows (1990) defined problem of detecting

ESP in reflected light for HST• Their conclusion was that it was unlikely

– Faint Object Camera incorporated an instrument designed for the direct detection of planets

• Major Project Scientist GTO Program: – A nearby star survey for ESPs- Faint Companions

• Bill fastie (JHU) & Dan Schroeder (Beloit)

• Key HST contributions have so far come from indirect observations: transit spectroscopy and astrometry

Mark Clampin/GSFC

Example Programs – Cycles 1 - 10

• TMR-1C – Tereby et al. (2000)– Initially believed to be 2-

3 MJ planet ejected by binary system

– Now believed to be a background star

• Schroeder et al. (2000)– Search for faint

companions with WFPC2– Adjusted to focus on

Brown Dwarf survey– Surveyed 23 stars within

13 pc using 1042M filter

TMR-1C, Tereby et al 2000)

Proxima Cen (Schroeder et al. 2000)

Mark Clampin/GSFC

Requirements: Summary

• Detection of planets in reflected light– Brown & Burrows (1990)– Jupiter analogs require contrast ratios ~109

– Terrestrial planet analogs require contrast ratios ~1010

• Jupiter analog @ 10 pc– 10-8 contrast ratio @ 0.16”

• 1.6 Au star-ESP separation, m = 20 mag

– 10-9 contrast ratio @ 0.5”• 5 Au star-ESP separation, m = 22.5 mag

– 10-10 contrast ratio @ 1.6”• 16 Au star-ESP separation, m = 25 mag

Mark Clampin/GSFC

Requirements

From Burrows, Sudarsky, and Hubeny 2004, Astroph 0401522

Mark Clampin/GSFC

What is a Coronagraph ?

Lyot stop

Fieldstop

Entrance pupil

Finalimage

IncidentStarlight

Field angle (arcsec)

Fie

ld a

ngle

(ar

csec

)

Intens after field occulter

-0.1 -0.05 0 0.05 0.1

-0.1

-0.05

0

0.05

0.1

Pupil X (meters)

Pup

il Y

(m

eter

s)

log10 Intens before Lyot stop

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

Pupil X (meters)

Pup

il Y

(m

eter

s)

log10 Intens after Lyot stop

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

Field angle (arcsec)

Fie

ld a

ngle

(ar

csec

)

log10 intens at image

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Broadband speckle pattern Speckle pattern

Courtesy

Mark Clampin/GSFC

Past/Current HST Capabilities

• Faint Object Camera– f/288 channel with coronagraph – best design– Never used in as-designed configuration– Photon counting precluded bright target observations

• STIS – Partial coronagraph in direct imaging mode– Two intersection occulting bars– Bandpass defined by CCD detector (no filters)

• NICMOS– Camera 2 implementation– Occulter is a hole in mirror

• Advanced Camera for Surveys– High resolution camera: fully sampled PSF– Aberrated beam design

Mark Clampin/GSFC

HST Optics

• Wavefront control is the primary issue in determining coronagraph performance

• Light is coherently diffracted by the residual polishing errors in the optics– Structure at n cycles/aperture

diffracts light to n Airy radii – Mid-frequency errors 5-50 cpa

are key issues for planet detection

– HST WFE~18 nm mid-freq.

– Apodizing masks to control diffraction have not always been sized for optimum coronagraphic performance on HST instruments

• HST telescope is subject to thermal variation (breathing)

Mark Clampin/GSFC

Prospects with ACS

• The high resolution camera in the Advanced Camera for Surveys is the most recent addition to HST’s instrument complement

• ACS employs an “aberrated beam” coronagraph– Occulting mask is in uncorrected focal plane– Fully-sampled PSF in visible– Best performance of any operational HST

coronagraph– Trade we make for performance is limited inner

working angle– Two occulting masks 1.8”, and 3.0” diameter masks

Mark Clampin/GSFC

ACS Coronagraph

Mark Clampin/GSFC

HRC-Coronagraph Performance

1.8" Spot Azimuthal Median Profiles

0 2 4 6 8Arcsec

10-10

10-8

10-6

10-4

10-2Flu

x A

rcse

c-2 (re

lativ

e to

tota

l ste

llar

flux)

Coronagraph only

Coronagraph - star

Direct (no coronagraph)

F435WF814W

Mark Clampin/GSFC

HD141569A – Circumstellar Disk

ACS – Clampin et al. 2003

NICMOS – Weinberger et al. 1999

Mark Clampin/GSFC

The challenge!

Mark Clampin/GSFC

Spectral Deconvolution

• GTO team have a program in progress to search for planets around Cen– Initially planned straightforward imaging program

• ACS Coronagraph does not achieve requirements for planet detection even with PSF subtraction

• Employ technique developed by Sparks & Ford (2003)– Spectral deconvolution– ACS lends itself to this technique since it has a

complement of ramp filters offering narrow and medium band imaging.

Mark Clampin/GSFC

Spectral deconvolution

Wavelength

Sparks & Ford 2002 ApJ 578, 543

Mark Clampin/GSFC

Narrow band coronagraphic observations HD130948: FR656N

634 nm

648 nm

663 nm

678 nm

R=60

Mark Clampin/GSFC

Narrow band coronagraphic observations rescaled

Mark Clampin/GSFC

Processed FR656N observation

Mark Clampin/GSFC

Medium band coronagraphic observations HD130948: FR914M

791 nm 863 nm 940 nm1025 nm

R16

Mark Clampin/GSFC

Processed FR914M observation

Mark Clampin/GSFC

FR914M 8-point performance

5 detection W of HD130948 (artificial stars at cardinal points)

Mark Clampin/GSFC

Medium band coronagraphic observations HD130948: FR914M

Spectral information:L-dwarf companionCount rates:791 nm 11.6863 nm 124.940 nm 127.1025 nm 42.

Mark Clampin/GSFC

ACS Medium Band (FR914M) Coronagraphic Observations of HD130948: “Spectral Deconvolution”

200 sec @ 791 nm

The primary is a G2V at a distance of 17.9 pc, V 6.1

200 sec @ 863 nm200 sec @ 940 nm200 sec @ 1025 nm

Spectral Deconvolution of the ACS Medium Band Coronagraphic Observations HD130948

The L-dwarf binary is 2.64 from the primary. The binary separation is 0.134. The binary is ~ 13 mags fainter than the primary in these filters.

Sparks, W. & Ford, H., “Imaging Spectroscopy for Extrasolar Planet Detection” 2002 ApJ, 578, 543

We get within square root two of the shot noise in the speckles!

This should allow detection of a Jupiter around Cen A or Cen B.

Mark Clampin/GSFC

Spectral deconvolution implementation parameters

Spectral Resolution

Maps onto “outer working distance” . Speckles smear if they move by /D, dark space fills in. Can still use.If is bandpass (one spectral resolution element) and 1+(/D) then it follows that spectral resolution R / = /(/D) = number of Airy rings

Wavelength Range

Maps onto “inner working distance”. A speckle has to move by ~ (/D) so for an inner working distance of N Airy rings at wavelength , I.e. N= /(/D), at wavelength there are N = / (2/D) rings. For N1-N2=1 require 2 =(N1/N ) =(1+N2)/N .

Mark Clampin/GSFC

Spectral deconvolution advantages

Improves detection process by eliminating speckles (in data analysis: does not eliminate their shot noise). Offers detection at small values of Q; recognition of speckles. Provides robustness against systematics.

Begin characterization from the outset since obtaining spectrum is implicit part of process.

Maximum observing efficiency because of spectral multiplex. Automatically observe all candidates in field; obtain detection and characterization in single observation.

Requires narrow or medium band imaging

Next:Take advantage of dark space between speckles to improve beyond average photon detection limit

Mark Clampin/GSFC

Spectral deconvolution

Subtracted PSF reveals extrasolar planet at S/N=20 with Q=0.01

(Sparks & Ford 2002)

Mark Clampin/GSFC

Cen: preliminary results

W. Sparks and R. White & ACS Science Team

•ACS offers opportunity to apply this technique to a very small number of nearby stars where the separation and m are favorable

• Cen is the most obvious candidate

Mark Clampin/GSFC

Cen: preliminary results

W. Sparks and R. White & ACS Science Team

• Preliminary results showing Lucy deconvolution performed on ~10% of data set

• Right hand image shows example with different bands & weighted sum

Mark Clampin/GSFC

STIS

• STIS– Definitive summary of capabilities in Grady et al. (2003) – Does not achieve contrast levels of ACS, closer inner

working angle – defined by two occulting bars

Mark Clampin/GSFC

Prospects with NICMOS

• NICMOS coronagraph has been very effective in discovering debris disks and faint companions

• NICMOS offers the opportunity to pursue searches for very young “self-luminous” planets which might be detectable in the near-IR– This is a key focus of

ground-based AO programs– NICMOS offers a very stable

platform with superior stability for such programs

– NICMOS has the benefit that it can observe closer to central star than ACS

TWA6 Field: h=13.2 @ 2.5” (Schneider 2002)

HD 141569 (Weinberger et al. 1999)

Mark Clampin/GSFC

NICMOS Capabilities

Discovery Space0

6 7 8 9 10Log10 Age (years)

80Mjup

14Mjup

JUPITER

SATURN

STARS (Hydrogen burning)

BROWN DWARFS (Deuterium burning)

PLANETS

200Mjup

Evolution of M Dwarf Stars, Brown Dwarfsand Giant Planets (from Adam Burrows)

-10

-8

-6

-4

-2

Log

L/L

(sun

)

Mark Clampin/GSFC

NICMOS Discovery Space

• Key resource:– Domains of observability in the near-IR with

HST/NICMOS and (Adaptive Optics Augmented) Large Ground-Based telescopes – Schneider 2000.

• Planet Detection capability– 10 Myr can detect 1 Mj at r=2”– 1 Myr can detect 1 Mj at r=1”– Bias towards planets at larger distances from star

• Cycle 13 saw large number of coronagraph programs awarded

Mark Clampin/GSFC

Cycle 13 Survey• Major program in Cycle 13 by

Inseok Song et al.

• Survey of 116 young nearby stars• • The selected targets are young (<~

50 Myr) and nearby (<~55 pc)

Detectable minimum mass planets

Mark Clampin/GSFC

Summary

• Key Science still to be done– With its current complement of instruments HST

offers the possibility to:• Detect young-planets with NICMOS

– Could apply spectral deconvolution to NICMOS too• Detect planets around a few stars with ACS if spectral

deconvolution can be fully exploited

• Issues– Both these resources require HST’s current pointing

capability– Coronagraph observations with HST carry a

significant overhead to perform observations effectively

• Require multiple rolls & comparison stars– Significant allocations of time are required to fully

exploit the ACS & NICMOS capabilities for detection of extra-solar planets

Mark Clampin/GSFC

New Instrumentation

• HST would have an instrument optimized for detection of planetary systems– Brown et al. proposed such an instrument CODEX– CODEX employs a deformable mirror to correct mid-

frequency WFE

• Alternative approaches are also possible– Phase and amplitude correction

• Labeyrie (2001)• Bowers and Woodgate (2003)

– Visible nulling coronagraph (Shao)

Mark Clampin/GSFC

Acknowledgements

• Thanks to Bill Sparks, Rick White, John Krist and Inseok Song.

Mark Clampin/GSFC